Executive Briefings

In a Complex Retail World, Kellogg Company Turns to Demand Sensing

The big name in cereal and snacks seeks to improve forecast accuracy through the adoption of a new tool that supplements traditional supply-chain planning processes with a broader range of information about actual customer demand.

Life in the cereal aisle of the grocery store used to be so simple. There were the familiar brands, popular variations on oats, corn, wheat and rice, and they rarely changed. Today, that sense of continuity has vanished. New products pop up at a rapid rate, and others disappear, as merchandisers struggle to get the attention of jaded consumers. And with this dizzying array of choices comes the challenge of guessing what's going to sell.

Predictability, it seems, has gone out the window. So companies are turning to sophisticated demand-sensing tools to get a better notion of actual buying patterns. Such was the motive behind The Kellogg Company's acquisition of a demand-sensing application from Terra Technology.

Posting sales of more than $13bn in 2011, Kellogg is the world's leader maker of cereal products. It's also a major player in snacks and convenience foods. Brands include Keebler, Pop Tarts, Eggo, Cheez-It, Famous Amos, Morningside Farms and Pringles, the last of which it recently acquired from Procter & Gamble.

According to Kurt Specht, who oversees demand planning for Kellogg USA, the company's need for a better system was driven by changing business dynamics, "due primarily to trade promotion activity, coupled with a strong innovation focus."

Like any successful producer in the consumer-packaged goods arena, Kellogg can't afford to stand still.

Traditionally, Kellogg's demand-planning and forecasting resources have been aimed at the intermediate term. Most CPG companies, says Specht, tend to focus on one to three months in the future. Less attention is paid to the very short or very long term.

"Often if you're looking at next week or the week after, a lot of your key decision-making on raw materials production planning is pretty well set already," says Specht. "What we came to realize in talking to Terra was that, by utilizing a capability such as theirs in the short term, there was an additional opportunity to improve forecast accuracy and address supply-chain efficiencies."

The 'Cascading' Effect

A company that can improve the forecast for next week or the week after creates a "cascading effort" that affects production planning for the future as well, says Specht. "The quality and accuracy of your call is better in the near term, which means that inventory is more appropriate for that plan. You are able to influence production quality and timing."

Better demand sensing comes to the fore as product lifecycles shrink, and the pace of new-product introductions picks up. According to an annual benchmarking study by Terra Technology, just 51 percent of CPG items have more than two years of history today. "Half of them you can't fit a [traditional] forecasting model to, but you still need to forecast," says Terra chief executive officer Rob Byrne. "Demand sensing is one alternative."

The method serves as an alternative to traditional demand planning, whereby companies look primarily at seasonal patterns and average sales for a given time of year. Demand sensing, says Byrne, "brings in all the data that might be relevant about what's going on in the marketplace." Orders, historical trends, point-of-sale data and channel inventory are among the inputs that can be factored into the mix.

As a software provider, Terra was on Kellogg's short list of candidates early on. Richard Dregne, vice president of sales and operations planning with Kellogg North America, said the company was attracted by the vendor's technology as well as its client base, "which was very targeted. They demonstrated to us that they had been successful with companies similar to our own. We didn't really find anybody else that had the breadth."

Implementation at Kellogg began with a pilot in November of 2009, with a focus on the U.S. and Canada. "We provided a large subset of our data to Terra, and worked with them to simulate their software as it would exist in production, to understand just what they could do for us," says Dregne. "It was an extremely important and compelling part of the value proposition, helping us be able to go forward."

Forecast Accuracy Improves

Benefits were quickly evident. For short-term demand over two weeks, Kellogg experienced 40-percent greater accuracy with Terra's technology, Dregne says.

As of early June, Kellogg was halfway through its implementation of demand sensing, and had yet to go live with any of its product lines. The company had encountered some challenges related to defining specific data requirements and ensuring consistency across product lines and channels. "Thus far," says Dregne, "the process has been fairly smooth, as we anticipated the hurdles and planned for them." Kellogg is targeting later this year as the date for full implementation, he adds.

The company also plans to evaluate whether to extend the application internationally. As for drawing on additional inputs, "we next plan to focus on consumer decision-making, using both our internal data and retailer data."

That's a transition that many CPG companies will soon be making. Introduced in 2002, the earliest version of Terra's demand-sensing tool wasn't set up to handle point-of-sale data, Byrne says. The capability was added with introduction of the vendor's multi-enterprise version, although many companies have yet to take full advantage of it. They've been occupied with accessing internal data, often contained in their enterprise resource planning applications, to supplement sales-trend patterns.

Suppliers and buyers are just beginning to share retailer data in their efforts to improve the forecast, but much of the activity right now is around the exchange of "transient, short-term signals," says Byrne. The ultimate goal between partners, he adds, is collaborative replenishment optimization, allowing merchandisers more precisely to match supply with demand.

As for Kellogg, it's bullish on the prospects of improved forecasting through use of the Terra model. "We are confident that the remainder of the project will go well," says Dregne. "We anticipate that this will help us to deliver much better results from forecast accuracy in the near term, which will allow us to provide better customer service and better control our supply-chain costs."

Resource Links:

Kellogg
Terra Technology


Keywords: supply chain, supply chain planning, retail supply chain, supply chain forecasting, Kellogg Company, Terra Technology, consumer packaged goods, grocery industry, retail demand-sensing

Life in the cereal aisle of the grocery store used to be so simple. There were the familiar brands, popular variations on oats, corn, wheat and rice, and they rarely changed. Today, that sense of continuity has vanished. New products pop up at a rapid rate, and others disappear, as merchandisers struggle to get the attention of jaded consumers. And with this dizzying array of choices comes the challenge of guessing what's going to sell.

Predictability, it seems, has gone out the window. So companies are turning to sophisticated demand-sensing tools to get a better notion of actual buying patterns. Such was the motive behind The Kellogg Company's acquisition of a demand-sensing application from Terra Technology.

Posting sales of more than $13bn in 2011, Kellogg is the world's leader maker of cereal products. It's also a major player in snacks and convenience foods. Brands include Keebler, Pop Tarts, Eggo, Cheez-It, Famous Amos, Morningside Farms and Pringles, the last of which it recently acquired from Procter & Gamble.

According to Kurt Specht, who oversees demand planning for Kellogg USA, the company's need for a better system was driven by changing business dynamics, "due primarily to trade promotion activity, coupled with a strong innovation focus."

Like any successful producer in the consumer-packaged goods arena, Kellogg can't afford to stand still.

Traditionally, Kellogg's demand-planning and forecasting resources have been aimed at the intermediate term. Most CPG companies, says Specht, tend to focus on one to three months in the future. Less attention is paid to the very short or very long term.

"Often if you're looking at next week or the week after, a lot of your key decision-making on raw materials production planning is pretty well set already," says Specht. "What we came to realize in talking to Terra was that, by utilizing a capability such as theirs in the short term, there was an additional opportunity to improve forecast accuracy and address supply-chain efficiencies."

The 'Cascading' Effect

A company that can improve the forecast for next week or the week after creates a "cascading effort" that affects production planning for the future as well, says Specht. "The quality and accuracy of your call is better in the near term, which means that inventory is more appropriate for that plan. You are able to influence production quality and timing."

Better demand sensing comes to the fore as product lifecycles shrink, and the pace of new-product introductions picks up. According to an annual benchmarking study by Terra Technology, just 51 percent of CPG items have more than two years of history today. "Half of them you can't fit a [traditional] forecasting model to, but you still need to forecast," says Terra chief executive officer Rob Byrne. "Demand sensing is one alternative."

The method serves as an alternative to traditional demand planning, whereby companies look primarily at seasonal patterns and average sales for a given time of year. Demand sensing, says Byrne, "brings in all the data that might be relevant about what's going on in the marketplace." Orders, historical trends, point-of-sale data and channel inventory are among the inputs that can be factored into the mix.

As a software provider, Terra was on Kellogg's short list of candidates early on. Richard Dregne, vice president of sales and operations planning with Kellogg North America, said the company was attracted by the vendor's technology as well as its client base, "which was very targeted. They demonstrated to us that they had been successful with companies similar to our own. We didn't really find anybody else that had the breadth."

Implementation at Kellogg began with a pilot in November of 2009, with a focus on the U.S. and Canada. "We provided a large subset of our data to Terra, and worked with them to simulate their software as it would exist in production, to understand just what they could do for us," says Dregne. "It was an extremely important and compelling part of the value proposition, helping us be able to go forward."

Forecast Accuracy Improves

Benefits were quickly evident. For short-term demand over two weeks, Kellogg experienced 40-percent greater accuracy with Terra's technology, Dregne says.

As of early June, Kellogg was halfway through its implementation of demand sensing, and had yet to go live with any of its product lines. The company had encountered some challenges related to defining specific data requirements and ensuring consistency across product lines and channels. "Thus far," says Dregne, "the process has been fairly smooth, as we anticipated the hurdles and planned for them." Kellogg is targeting later this year as the date for full implementation, he adds.

The company also plans to evaluate whether to extend the application internationally. As for drawing on additional inputs, "we next plan to focus on consumer decision-making, using both our internal data and retailer data."

That's a transition that many CPG companies will soon be making. Introduced in 2002, the earliest version of Terra's demand-sensing tool wasn't set up to handle point-of-sale data, Byrne says. The capability was added with introduction of the vendor's multi-enterprise version, although many companies have yet to take full advantage of it. They've been occupied with accessing internal data, often contained in their enterprise resource planning applications, to supplement sales-trend patterns.

Suppliers and buyers are just beginning to share retailer data in their efforts to improve the forecast, but much of the activity right now is around the exchange of "transient, short-term signals," says Byrne. The ultimate goal between partners, he adds, is collaborative replenishment optimization, allowing merchandisers more precisely to match supply with demand.

As for Kellogg, it's bullish on the prospects of improved forecasting through use of the Terra model. "We are confident that the remainder of the project will go well," says Dregne. "We anticipate that this will help us to deliver much better results from forecast accuracy in the near term, which will allow us to provide better customer service and better control our supply-chain costs."

Resource Links:

Kellogg
Terra Technology


Keywords: supply chain, supply chain planning, retail supply chain, supply chain forecasting, Kellogg Company, Terra Technology, consumer packaged goods, grocery industry, retail demand-sensing